A Human-Centered, Reinforcement Learning-Driven Robotic Framework for Collaboratively Supporting On-Site Construction Workers

In preparation, doctoral dissertation.

Yuning Wu

Committee: Daniel Cardoso Llach, Jean Oh, Jieliang Luo
I am currently in Ph.D. ABD (All But Dissertation) stage.

Towards Human-Centered Construction Robotics: An RL-Driven Companion Robot For Contextually Assisting Construction Workers

Under review, submitted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).

Yuning Wu, Jiaying Wei, Jean Oh, Daniel Cardoso Llach

/ PDF / arXiv / video

TL;DR: This paper documents the technical development of a socially-aware "work companion robot" focused on assisting, rather than brutally replacing, skilled human workers. By taking on the more physically burdensome tasks within an existing prototypical construction workflow, specifically carpentry formwork, the robot enables workers to concentrate on the more high-skilled and sophisticated facets of their daily practice. The paper provides a systematic overview of how the robot prototype is architected, driven by recent RL-based social navigation methods. It also proposes a novel pipeline for incrementally fine-tuning a pretrained RL model to align with context-specific worker behaviors while preventing drastic policy shifts in Sim2Real. The robot is evaluated on an actual construction site by a group of workers through direct operation.

Construction as Computer-Supported Cooperative Work: Designing a Robot for Assisting Carpentry Workers

In revision, submitted to ACM Conference on Computer Supported Cooperative Work (CSCW).

Yuning Wu*, Emek Erdolu*, Jiaying Wei, Jean Oh, Daniel Cardoso Llach

TL;DR: This paper reimagines the integration of robotics into highly manual construction work from a socio-technical perspective. It emphasizes the application of Computer-Supported Cooperative Work (CSCW) principles, presenting a case study on the design, development, and evaluation of a prototype robot designed to assist in an existing indoor construction workflow, carpentry formwork. The paper discusses the use of ethnographic and practice-based research methods to understand the cooperative work of construction teams and to inform the design of supportive robot technologies. It proposes specific human-centered benchmarking metrics for evaluating the envisioned robot support and advocates for an approach that supports, rather than completely reconfigures or replaces, the human elements in existing labor-intensive work.

Learning Dense Reward with Temporal Variant Self-Supervision

[ICRA'22] IEEE International Conference on Robotics and Automation 2022 | RL for Contact-Rich Manipulation Workshop.

Yuning Wu, Jieliang Luo, Hui Li

/ PDF / arXiv / slides / code

TL;DR: This paper introduces an efficient framework for deriving dense rewards in reinforcement learning, aimed at enhancing the efficiency and robustness of training robots for contact-rich manipulation tasks. Developed through a collaboration between Carnegie Mellon University and Autodesk Research, this framework utilizes Temporal Variant Forward Sampling (TVFS) and a self-supervised learning architecture. By leveraging temporal variance and multimodal observation pairs, the generated dense reward facilitates faster and more stable policy training in a plug-and-play manner, exemplified in tasks such as joint assembly and door opening.

Towards a Distributed, Robotically-Assisted Construction Framework: Using Reinforcement Learning to Support Scalable Multi-Drone Construction in Dynamic Environments

[ACADIA'20] Annual Conference of the Association of Computer Aided Design in Architecture 2020.

Zhihao Fang, Yuning Wu, Ammar Hassonjee, Ardavan Bidgoli, Daniel Cardoso Llach

/ PDF / CumInCAD / slides / video / code

TL;DR: This paper explores a robotic framework that utilizes the multi-agent reinforcement learning method (MADDPG) to enable drones to execute collaborative tasks such as bricklaying and spray-coating, specifically designed for complex-shaped geometric structures. It showcases a robotic framework adept at managing intricate real-world designs in dynamic environments, leveraging recent advances in machine learning.

Teaching an AI Agent Spatial Sense by Playing LEGO

In preparation.

Yuning Wu, Katerina Fragkiadaki

TL;DR: Whether one is an architect or not, the way we design and build reflects our spatial understanding of the physical world. As children, this exploration typically begins by playing with building blocks. This research retraces these initial steps by training an AI agent to play with LEGOs in a physical simulator (NVIDIA Omniverse Isaac Gym / PyBullet). A crucial aspect of this study is enabling the AI agent to design valid, and eventually elegant, stacked structures within a specified arena by interpreting spatial relationships among the blocks. Various components of this research are pending completion and will be submitted to both top-tier AI and robotics conferences as well as an art venue.

Design Architecture with Graphs: Utilize Space Connectivity and Graph Neural Networks To Inform Layout Design

Master Thesis Project

Yuning Wu

/ link

TL;DR: This thesis project explores the potential of utilizing space connectivity and graph neural networks (GNNs) to inform layout design. It proposes a novel computational framework that leverages the spatial relationships between architectural elements to generate more informed and contextually relevant design solutions, contrasting from solutions that are purely data-driven or rule-based.

State Space Paradox of Computational Research in Creativity

[HI'19] Thirtieth Anniversary 'Heron Island' Conference on Computational and Cognitive Models of Creative Design.

Ömer Akin, Yuning Wu

/ PDF / arXiv / slides

In memory of Professor Ömer Akin (1942-2020). This is one of his last papers, and I am privileged to be organizing and presenting it. This paper explores the State Space Paradox (SSP) in computational creativity research, discussing how current digital systems, designed as closed systems with predefined parameters, are inherently limited in their creative capacity compared to the open systems of human creativity. It examines various procedural and representational approaches to artificial creativity, such as genetic algorithms and shape grammars, and critiques their ability to truly emulate human creativity. The paradox lies in the fact that while these systems can produce outputs that may initially seem creative, they are ultimately unable to exceed their predefined state spaces and thus fail to achieve genuine novelty or adapt beyond their initial programming.